|
Research Institute of Intelligent Computer Systems Ternopil National Economic University |
|
2006, Vol. 5, Issue 1 |
|
Contents and abstracts
FAST NEURAL NETWORK FOR GENERALIZED TRIGONOMETRIC TRANSFORMATIONS SYNTHESIS Oleh Liskevych 1), Mykhaylo Yatsymirskyy 2)
1) State Scientific and Research Institute of Information Infrastructure,
Troleibusna Str., 11, Lviv, 79053, Ukraine, e-mail: ol@rol.com.ua, www.dndiii.lviv.ua
The structure of the fast hardware neural network, based on generalized trigonometric transformations algorithm is developed. The network is appointed for optimal by some given criteria transformation selection and synthesis in adaptive digital signal processing system. COMMUNICATION BASED UPON ADVANCED RSS FEEDS CONCEPTS AND STRATEGIES Peter J. A. Reusch 1-2), Bastian Stoll 1), Pawel Serwatowski 1)
1) University of Applied Sciences - FH-Dortmund - Germany
RSS feeds provide information through channels that can be subscribed. Updates in subscribed channels can be controlled by feed readers automatically - to make new information available to the user - immediately after the update. Today elementary RSS feeds are widely used for information dissemination. Advanced RSS concepts can improve communication concepts in many areas. A set of RSS applications is described, new perspectives are mentioned, and strategies derived from the implemented applications. HIERARCHICAL DECOMPOSITION OF PETRI NETS FOR ANALYSIS AND DESIGN OF DIGITAL MICROSYSTEMS Andrei Karatkevich, Grzegorz Andrzejewski
University of Zielona Gora, ul. Podgorna 50, 65-246 Zielona Gora, POLAND.
The paper is dedicated to hierarchical Petri nets with application to design of control devices. It consists of two parts. The first part introduces a concept of simplifying analysis of flat Petri nets (we discuss analysis of reachability and liveness) by means of hierarchical decomposition. We present a class of subnets which can be replaced by the macroplaces. Further we describe properties of such decomposition and show, that analysis of the mentioned properties of a net can be reduced to the corresponding analysis of the elements of decomposition. A method allowing to obtain the decomposition is presented. The second part presents an approach to specification of a digital control system behavior by means of hierarchical Petri nets. We describe a model, called HPN, which has some of important properties, simplifying description of strongly reactive systems. SOFTWARE IMPLEMENTED HARDWARE-TRANSIENT FAULTS DETECTION Goutam Kumar Saha
Centre for Development of Advanced Computing, Kolkata, India
This paper examines a software implemented self-checking technique that is capable of detecting processorregisters' hardware-transient faults. The proposed approach is intended to detect run-time transient bit-errors in memory and processor status register. Error correction is not considered here. However, this low-cost approach is intended to be adopted in commodity systems that use ordinary off-the-shelf microprocessors, for the purpose of operational faults detection towards gaining fail-safe kind of fault tolerant system. METHOD OF HYSTERESIS CALCULATION IN A HYSTERESIS CONTROL CURRENT REGULATOR WITH CURRENT REFERENCE ADAPTATION Stanislav Koschinsky, Dmitry Tey
Radioelectronics Department, State Technical University of Orel, 40, Naugorskoye Shosse, 302020 Orel, Russia
In this paper hysteresis adaptation method is presented. It provides the algorithm to calculate hysteresis value equal to the maximum pulsation of current. This allows decrease disadvantage of hysteresis regulator with clocked commutation in application with variable reference. NONLINEAR PARAMETER IDENTIFICATION OF RISK TECHNICAL SYSTEMS Nikolay Petrov
Trakia University - Stara Zagora, Yambol, Bulgaria;
This paper deals with an optimal method concerning nonlinear parameter identification of risk technical systems (automobile and railway transport, aircrafts, marine and river transport, chemical installations, munitions, information society suffering by terrorism). Unknown states of the model are built by sliding observers which converge in a finite time. Due to this property, it is possible to derive equations of the model in order to obtain an estimation law which converges to the nominal values of the parameters also in the finite time. APPLICATION OF THE GROUP METHOD OF DATA HANDLING IN SYNTHESIS OF SUPRVISED NEURAL NETWORKS Homich Andrew Vladimirovich 1), Zhukov Leonid Alexandrovich 2)
1) Siberian state university of technology, Krasnoyarsk, prospectus of the World 82, 660049, Russia, supremum76@rambler.ru
The expediency of application of group method of data handling (GMDH) in synthesis supervised artificial neural networks is considered. Comparison of efficiency of two variants GMDH is carried out. Methods using entrance variables on everyone and only the first line of selection are considered. Efficiency of offered methods is estimated experimentally on practical problems. NEURO-FUZZY MODELLING IN ANAEROBIC WASTEWATER TREATMENT FOR PREDICTION AND CONTROL Snejana Yordanova 1), Rusanka Petrova 1), Nelly Noykova 2), Plamen Tzvetkov 1)
1) Faculty of Automation, Technical University of Sofia, 8 Kl.Ohridski blvd., 1000 Sofia, Bulgaria,
e-mail: sty@tu-sofia.bg
The aim of the present paper is to develop neuro-fuzzy prediction models in MATLAB environment of the anaerobic organic digestion process in wastewater treatment from laboratory and simulated experiments accounting for the variable organic load, ambient influence and microorganisms state. The main contributions are determination of significant model parameters via graphical sensitivity analysis, simulation experimentation, design and study of two “black-box” models for the biogas production rate, based on classical feedforward backpropagation and Sugeno fuzzy logic neural networks respectively. The models application is demonstrated in process predictive control. MIXING SPEECH SIGNALS FOR THE REAL-TIME MULTIMEDIA SYSTEMS A. Melnyk 1), T. Korkishko 2), R. Shevchuk 3)
1) Computer Engineering Department, Lviv Polytechnic National University, S. Bandery Str., 12, 79046,
Lviv, Ukraine, E-mail: aomelnyk@polynet.lviv.ua
In this work we considered basic principle of mixing and offered the method of multistage mixing which allows the mixing speech samples on the measure of receipt of data blocks in mixer, and also mixing speech samples, that was got from decompression of compression speech signals of different formats. DISCRETE REPRESENTATION MODEL OF MULTIPLY CONNECTED PIPELINES WITH DISTRIBUTED PHYSICAL PROPERTIES Igor Lyutak
Ivano-Frankivsk National Technical University of Oil and Gas, department of software of computer-based systems,
In this paper presented approach of developing chain model which describe distributed physical properties of multiply connected pipeline. It is shown that the digital autoregressive filter can be used to analyze signals of wave propagation in the chain model. A NEW VECTOR QUANTIZATION APPROACH FOR DISCRETE HMM SPEECH RECOGNITION SYSTEM Mohamed Debyeche 1), Jean Paul Haton 2) and Amrane Houacine 3)
1) Faculty of Electronics and Computer Sciences, USTHB, P.O.Box 32 El-Alia, Bab-Ezzouar, Algiers, Algeria,
e-mail: mdebyeche@caramail.com and mdebyeche@usthb.dz
In order to address accuracy issues of discrete Hidden Markov Models (HMMs), in this paper, a new vector quantization (VQ) approach is presented. This new VQ approach performs an optimal distribution of VQ codebook components on HMM states. This technique that we named the distributed vector quantization (DVQ) of hidden Markov models, succeeds in unifying acoustic micro-structure and phonetic macro-structure, when the estimation of HMM parameters is performed. The DVQ technique is implemented through two variants. The first variant uses the K-means algorithm (K-means-DVQ) to optimize the VQ, while the second variant exploits the benefits of the classification behavior of neural networks (NN-DVQ) for the same purpose. The proposed variants are compared with the HMMbased baseline system by experiments of specific Arabic consonants recognition. The results show that the distributed vector quantization technique increase the performance of the discrete HMM system. WEB ACCESSIBLE MULTIACCOUNT CONFIGURABLE CLIENT EMAIL APPLICATION Krzysztof Ratecki, Bartosz Sakowicz, Marcin Wojtowski, Andrzej Napieralski
Department of Microelectronics and Computer Science,
al. Politechniki 11, 90-924 Lodz, POLAND
The article presents alternative ways of checking emails and how to make an application mail enabled using library of J2EE Platform – JavaMail API. The article describes application attainable from WEB browser based on JSP Scripts and Java servlets. Application allows for access to multiple email accounts through one web page. TESTING AND DIAGNOSIS OF DISTRIBUTED DIGITAL SYSTEMS Vladimir Hahanov, Vladimir Yeliseev, Alexander Parfenty Kharkov National University of Radio Electronics, 14 ave Lenin, 61166 Kharkiv, Ukraine hahanov@kture.kharkov.ua The complex system can be represented as hierarchy of simpler components. On the one hand different levels of hierarchy have an advantage for more deeply research of properties of the complex system. On the other hand a flexible model and method of testing and diagnosing of the complex system need to be developed. Moreover, international IEEE testability standards should be applied.The research presented in the current paper offers approach of testing of complex digital system based on hierarchy scaling during diagnosis experiment. Several models of testing are proposed. Main principles of testing system organization are given. The result of the work is significant time reduction of test and diagnostic of system overall. CONSISTENCY OF DISTRIBUTED SYSTEM WITH ACTIVE INITIATOR PROCESS WITHOUT USELESS CHECKPOINTS Gopalan N.P., Nagarajan K.
Dept. of Computer Science and Engg, National Institute of Tech., Tiruchirappalli, Tamilnadu, India – 620 015
Checkpointing mechanism is the one of the best attractive approach for providing software fault tolerance in distributed message passing systems. This paper aims to implement a distributed checkpointing technique, which eliminates the drawbacks of the centralized approach like “domino effect”, “useless checkpoint” (checkpoints that do not contribute to global consistency), and “hidden and zigzag” dependencies. The proposed checkpointing protocol has a checkpoint initiator, but, coordination among the local checkpoints is done in a distributed fashion. This guaranty that no message would be lost in case of failure occurs, has been maintained in this work by exchange of information among the processes. However, there is no central checkpoint initiator, but each of the processes takes turn to act as an initiator. Processes take local checkpoints only after being notified by the initiator. The processes synchronize their activities of the current checkpointing interval before finally committing their checkpoints. Thus, the checkpointing pattern described in this paper takes only those checkpoints that will contribute to the consistent global snapshot thereby eliminating the number of useless checkpoints. THERMOELECTRICAL PROCESSES SIMULATION IN RESISTIVE SUPERCONDUCTING FAULT CURRENT LIMITER Dmytro Fedasyuk, Pavlo Serdyuk
Lviv Polytechnic National University , S. Bandery Str., 12, Lviv, 79013, UKRAINE
Simulations of superconducting fault current limiter (SFCL) include modelling of thermal and electrical behaviour of the limiter. In this paper the mathematical model for the analysis of thermoelectrical processes in resistive superconducting fault current limiter for microelectronics devices protection is presented. An effective and fast halfanalytical computer simulation of the model is presented. Simulation results are presented. MATERIAL FLOW ANALYSIS AND PLANT LAY-OUT OPTIMIZATION OF A MANUFACTURING SYSTEM Francesco Longo, Giovanni Mirabelli, Enrico Papoff
Mechanical Department, University of Calabria, Italy, 87036, Via P. Bucci, Rende (CS)
The paper proposes the results of a research project developed by the authors in collaboration with a production system working in the field of manufacturing wood products. The first step of the project was an accurate analysis of the system and the design of new production documents in order to collect data about the system itself. The collected data have been used to propose an initial solution of plant lay-out. The second step of the project was the construction of a simulation model. The model has been used to find an optimal plant- layout configuration by means of genetic algorithms with the goal of material handling cost and shop order flow time minimization. REGION-BASED CLASSIFICATION OF ELECTRO-OPTICAL IMAGES WITH NEURAL NETWORKS AND FUZZY LOGIC Iryna Petrosyuk, Yuri Zaichenko
National Technical University of Ukraine "KPI", Kyiv, Ukraine
This paper reports on a novel approach to the optical information processing for the hyperspectral remote sensing systems by means of developed unification algorithm of the two mathematical tools: the fuzzy logic and the neural network. New neuro-fuzzy classification algorithm for hyperspectral remote sensed images has been proposed. It is able to replace complicated empirical formulae, which require the knowledge of dependences of many input parameters that rapidly change during of range time and difficult for crisp determination. |